Kyle Richardson

AstaBench: Rigorous Benchmarking of AI Agents with a Scientific Research Suite

Analytica: Soft Propositional Reasoning for Robust and Scalable LLM-Driven Analysis

ArtifactLinker: Linking Scientific Artifacts for Automatic SOTA Discovery

Language Modeling by Language Models

Declarative characterizations of direct preference alignment algorithms

ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoning

Understanding the Logic of Direct Preference Alignment through Logic